Test errors are different (represented differently in the JUnit XML
output from jest-junit) to test _failures_. Failures are tests
which ran and failed, errors are for tests/test suites which
did not even run because the test code itself didn't build
or didn't execute correctly.
jest-junit has an option to enable reporting of test suite errors,
but test-reporter then interprets these as successful tests.
Motivation: when using a matrix job, or more than one kind of tests in a same workflow, we can end up with multiple summaries at the same time. This will lead to multiple anchors and html id that will no longer be unique.
This prefix option allow to disambiguate those anchors ; and keep them functional.
Refactor PHPUnit support into separate phpunit-junit parser
Instead of modifying the Java JUnit parser, this creates a dedicated
PHPUnit parser that properly handles PHPUnit's nested testsuite elements.
This keeps the parsers cleanly separated and allows for future PHPUnit-
specific features.
Co-Authored-By: Matteo Beccati <matteo@beccati.com>
Co-Authored-By: Claude Code <noreply@anthropic.com>
Add workflow and fixtures to test the behavior of fail-on-error and
fail-on-empty parameters across different scenarios:
- Passing tests with fail-on-error true/false
- Failing tests with fail-on-error true/false
- Empty results with fail-on-empty true/false
Scenario 4 (failing tests + fail-on-error=false) is expected to fail
until issue #217 is fixed, documenting the bug where check conclusion
shows 'success' even when tests fail.
The workflow outputs a GitHub Actions summary with a markdown table
showing all test results.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude <noreply@anthropic.com>
Add python-xunit-parser.ts with associated case statement
Add python-xunit to reporter docs in action.yml
Add tests
Update README
Resolves#244Resolves#633
